46 - iassisl quarterly

Providing Local

Data Services

by R. de Vries '

Steinmetz Archives/SWIDOC.

Amsterdam, The Netherlands

The Steimnetz Archive, as the Dutch national

data archive for the social sciences, is a

somewhat special case in the context of this

workshop. We function both as a data

clearinghouse and as a data library per se. As

data library, the archive operates in an area

that, in other countries, would be seen as a

"regional" (as opposed to "national"); the

following are some reasons why the Steiiraietz

can serve as an example of a data library

providing local data services.

dependency on someone else's computer

center. I.e., decisions on installation of

software packages, policy on how to

handle mass storage problems and the

safekeeping of magnetic tapes,

participation in networks, etc., are all

beyond our direct control; we can argue,

but have no real influence on such

decisions, let alone the means of

implementing them on our own.

dependency on the willingness of research

'Paper prepared for presentation at the

IFDO/IASSIST Conference. Workshop on Data
Services, Amsterdam, Mav 20 - 23, 1985.

organizations and their funding bodies to

deposit data (survey or otherwise) in the

Archive's holdings. Again, we have no

financial means with which to buy large

datasets, nor the manpower, in the case

of published statistical data for example,

to generate new data from published

sources.

a small stafT (5).

a strong emphasis on documentation and

reference service, aided by a reference

database containing study descriptions of

every stored dataset "Documentation"

here refers to the original questionnaire,

research report, print-outs of frequencies,

etc

given the national role of the Archive a

less than desirable situation, but in the

context of "local data servcies" quite

reasonable: we caimot give access to our

holdings through a network, nor is the

reference database available online. Data

exchange is via magnetic tape, and

available datasets are brought to the

attention of potential users via regular

newsletters and a published catalogue. A
service that is in my opinion typical of

local data services, data exchange on

floppy discs, is possible but we have

hardly any experience with it as yet

If one sees a data library as a "local" service

backed by a central data clearinghouse or

central acquisition and processing centre, then

indeed one expects an organisation with a small

StafT, using external computers and software,

concentrating on reference as well as actual

dissemination in as friendly a manner as

possible, generating subsets, documentation and

other special requests. Also, seen

geographically, the data librar>' should be within

one day's travelling distance for its users, for

consultation purposes. Given this defmiiion, the

Summer 1986



iassist quarterly - 47

Steinmetz archive does serve as an example of a

data library.

Are there any other local data services in

Holland for the social sciences? For survey

data: no. For statistical information at the level

of cities or regions: yes. There are several

specialised databases ovmed by government

organizations for planning and policy making,

and by university departments, for research and

training. These are, on the whole, "local" in

the sense of being accessible only to their own
people, not to outside researchers, either

academic or otherwise.

In the second part of this account, I will briefly

outiine our approach in the following areas:

data acquisition, data dissemination, storage and

maintenance of data, dociraientation and

reference services.

Data acquisition is accomplished by routinely

checking registers of ongoing research, social

science periodicals, and reports of finished

research. This is facilitated by the Steinmetz

Archive's participation in the Social Science

Information and Documentation Centre. There

is no human network of researchers or fund

raisers in the field who could report to the

Archive interesting projects or data. (Nor would

I would expect the kind of data library that

provides local data service to rely on such a

network for data acquisition.)

Dissemination. Users in Amsterdam, where the

Steinmetz Archive is located, have direct access

to the data through the local university-owned

computer centre (SARA). From a local

termimil, a user can get a copy of a datasei by

simply starting a job, that has only one

variable: the Steinmetz number given to the

particular dataset Central logging of these jobs

and who has started them, is automatically

reported to the Archive, thus enabling a

monthly overview of this type of usage. Other

users receive the data, and often an SPSS setup,

on tape. This arrangement makes, of course, no

provision for users without at least access to a

minicomputer with a tape drive and statistical

package, such as SF*SS. For example, we are

imable by these means to provide service to

schools. Data transfer to such users should be

through floppy discs, a service that we have not

really started yet

As an archive, with an obligation to disseminate

10 to 15 year old datasets, it requires that we
have strong Data storage and maintenance

systems. This we achieve with a system of

multiple tape backups and a tape refreshing

scheme to guarantee that no tape is physically

more than three or four years old. All tapes

are stored in the computer centre. One might

expect a data library per se to be more relaxed

in these matters; whatever gets lost can be

replaced upon request from a central data

organisation, but this is not so in our case.

Developments in "laser disc" technology could

ease local mass storage problems, and at the

same time ensure long term reliability.

How is the user introduced to these masses of

carefully preserved data? Through

Documentation and reference services. The
Steinmetz Archive, as mentioned previously,

helps users find data suitable to their needs

through a catalogue, which is easily, and

regularly, produced from a reference database,

various indices on microfiche and paper, and

through the original documentation produced by

the principal investigator. Introductions to the

principles of empirical research and the data

available for secondary analysis by means of

making "teaching packages" available and giving

lectures at schools and colleges, are other means

to assist users. The Steinmetz does not give

lectures but does offer a teaching package

together with the relevant data; the package was

developed by an outside institution. Data

libraries, which should need to put less effort

into such activities as acquisitions, processing

and maintenance, might profitably put more

efibrt into actively getting users acquainted with

computer-assisted analysis, data sources, etc. n

Summer 1986